back to article Feeding the XPoint cuckoo and finding it a place in the storage nest

XPoint memory is a spectrum-invading cuckoo. Spectrums are great ways to describe a range including different items with differing characteristics. Everyone gets used to the range, then something comes along and forces its way into the spectrum, making existing items shuffle up and down. There is a memory-storage spectrum with …

  1. Anonymous Coward
    Anonymous Coward

    Another use

    XPoint could potentially be the speedier component in hybrid drives, which NAND currently occupies. Since it is a smaller part of an hybrid drive then the cost increase won't be so bad and boot/wakeup times will be excellent. Of course this is aimed more at consumer devices than servers.

  2. Permidion


    SARA ?

  3. Tom_


    This could be very attractive to Sony, Nintendo, MS and maybe others by the time the next generation of consoles come out, around 2020-2022, as by then the tech will have settled and prices might not be too high. That's possibly wishful thinking, though.

    In fact, it might provide a very clear differentiator for a new player to enter the market, such as Apple or Amazon, with a console that will have a noticable advantage over the PS4 and XBone. ie. Much better loading times.

    1. Jim84

      Re: Consoles

      I think Nintendo's next console (the Nintendo NX) is coming out a lot sooner than 2020 given the failure of the WiiU. Although XPoint will probably be too expensive.

      Can someone explain how XPoint will lead to better games?

  4. This post has been deleted by its author

  5. Andrew Harding

    There is a lot of F.U.D. with TLC overwrites and that limits the uses cases. It's hard not to think Xpoint can't take advantage of it.

    A very high limit of overwrites on a 1TB DIMM with a server with 96 slots would be attractive on hyperconverged if they can get the proper erasure coding, inline dedupe/compression and ECC equivalent.

    And we thought 10Gb/s NICs would be enough :p

  6. Magellan

    3D XPoint is a new form of RAM, not SSD

    Back on June 11th, seven weeks before the Intel/Micron announcement, Stephen Breezy at Seeking Alpha pulled back the covers on the purpose of the then unnamed 3D XPoint memory. Breezy has been watching Intel patent applications, and saw one for a hybrid DRAM/PCM memory architecture, which looks exactly like the "Apache Pass" memory architecture for Intel's Skylake CPU/Purley platform, the details of which leaked in late May.

    1. Mark Honman

      Re: 3D XPoint is a new form of RAM, not SSD

      Yes - if anything it reminds me of the merits of magnetic core, without the poor density and high power consumption.

      The most interesting use, short-term, would be paging store. Systems already have a mechanism for paging stuff in and out of memory, but it's pretty useless these days as mechanical drives are too slow/thrashy and SSDs wear out.

      If you view the paging mechanism as providing a RAM cache of the contents of the backing store (as if it were just a giant binary file that has been memory-mapped), the application is pretty obvious.

      For HPC this means that the size of the data set can exceed the size of available RAM; and as RAM no longer limits the data set size, new RAM tradeoffs become possible. The amount of RAM can be chosen to suit the number of pages that are "hot" at any given time, thereby reducing cost/heat/size. The savings then become available for more processing power.

      There is one element that would be a handy addition to the present virtual-memory model: a prefetch capability (analogous to that used in floating-point DSPs) that would extrapolate data-access patterns to identify data that should be brought into higher levels of the memory hierarchy before it is needed.

      By "higher levels" I'm really thinking of processor caches rather than DRAM, which is pitifully slow in comparison to the performance of HPC compute engines.

    2. semimurphy

      Re: 3D XPoint is a new form of RAM, not SSD

      Intel/Micron has re-invented the capability of NOR Flash.

      XPoint endurance is 1,000,000 write cycles, just like flash P/E endurance around the year 2000. The read speed is almost as good as 15 year old NOR Flash. Write speed has not been disclosed. PCM type memories are relatively fast to write, about 100 nanoseconds or better, and I think it likely the memory cell plus switch is a thin film diode in series with a chalcogenide glass stack up of some sort, so a similar temperature based write time is likely.

      (The above assumes 1000X write endurance of the most fragile available nand flash endurance of 1,000 program/erase cycles and 1000th the ~70 usec access of 3D TLC nand flash or 70 ns read access times, compared to 55 ns access times for NOR Flash).

  7. Alan Brown Silver badge

    Sales opportunity

    If 1Tb of extra memory is cheaper than ram, then losing sales in database farms will be more than made up by sales in other arenas (researchers will leap at the opportunity for holding more stuff in (slower) ram over having to page to disk.)

    On the other hand, if the database farm can do the same with fewer systems, then it can do much more with the same number of systems. Many of the larger players will increase capacity/capabilities rather than cut down on hardware.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2022